I've been a desktop Linux user for seventeen years. For eight of those years, I haven't had a copy of Windows installed on any machine in the house.
Under the circumstances, I can easily drift into thinking that Linux and free software represent the whole of computing, and possibly its future as well. Increasingly, though, I find myself wondering how long desktop Linux will last, especially when I consider the growing popularity of mobile devices.
Among manufacturers, of course, free software is thriving. Because free software's licenses allow manufacturers to use existing software, they have become a short cut to market. What might take five years to develop by proprietary means can take less than two when free software is used.
As a result, free software is used in everything from the Raspberry Pi to super-computers. Increasingly, too, free software is powering the Internet of Things as well, enabling an increasing number smart devices ranging from household appliances to cars.
But on the desktop? The latest figures I can find for Linux use is about 1.5% -- an increase of less than half a percent in the last seven years. The figures can be debunked as being too American-centered, or as simply being biased against Linux. But, regardless of which figures you accept, desktop Linux use lags far behind that of OS X, let alone Windows. Linux remains an also-run, a distant third among operating systems. By contrast, Linux adoption on web clients is a healthy second with just under 29%, while on main frames the figure is almost 99%.
As many hopes as I have for free software, even I have to admit that the figures for desktop Linux fall short of the goal of world domination. As little as I like the idea, desktop Linux is becoming less of an active force, whose use can only decline.
If desktop Linux is becoming irrelevant, it is not alone. Fewer desktop computers in general are being sold. According to IDC, 11.2% fewer personal computers shipped in the first quarter of 2016 than during the same period in 2015. Gartner put the decline at only 9.6%, but no matter which figure you accept, the decline has been going on for several years.
The explanation for this decline is usually the popularity of mobile devices. Admittedly, tablets themselves had 5% fewer sales in 2015, but that decline is just over half of the decline in personal computers. Moreover, as the first tablets sold increasingly reach the end of their working life, tablet sales are predicted to increase in 2016.
These sales figures are natural enough. Both tablets and phones represent the kind of computing that many users need. Instead of buying an expensive, all-purpose computer, which was once the norm in computing, most users find that a slower, smaller computer is all they need for texting and using the web. Business and artistic users may still need a desktop computer for the storage, speed, and widescreens, but they are a minority and always have been. The only difference now is that the market no longer caters to them.
The transition to mobile devices indirectly affects free software because, in general, it has never moved beyond the personal computer.
Aside from releasing the GNU Affero Public License in 2007, free software has done little to address the challenges to software freedom represented by the cloud services that are so central mobile devices.
For example, while the developers of desktop environments like GNOME and KDE have discussed moving into mobile devices, and even designed software for them from time to time, next to nothing has actually been done. Similarly, Krita, the popular paint application, only got around to a mobile version a couple of years ago, because they lacked the necessary developers.
Even more importantly, while Android is a modification of Linux, parts of it are released under the Apache License 2.0, which allows the software it covers to be re-released under a proprietary license. This provision allows Android to be treated as proprietary by the large mobile manufacturers, while the change of name means that Linux and free software get no credit for Android.
True, there is CyanogenMod, a free-licensed version of Android, but it is even less known Linux. Although CyanogenMod-based phones and tablets exist, they remain a distinct minority, being used in less than a hundred different models at best.
In comparison, I know of only one Linux-based tablet on the market: BQ's recently released Aquaris M10 Ubuntu Edition. Due to market shifts and a slowness of response, free software has been has been almost completely shut out the areas in which most of today's computing is being done.
Under the circumstances, the wonder is not that Linux remains a minority operating system. The wonder is that it has managed to increase its use by even the tiny amount that it has.
Perhaps the increased concerns about security and privacy, and anti- Microsoft sentiment, explain the increases. Yet, if so, these issues have barely managed to stalemate market forces, even though they have been widely discussed in the media.
Under these circumstances, how long can Linux and free software remain on the desktop? Increasingly, developers may be pressured to stop working on desktop projects and to build what businesses rather than home users want. When the present handful of desktop Linux users fade away, perhaps desktop Linux, too, will end, having lasted a single generation.